Relationships Among Non-Acremonium Sp. Fungal Endophytes in Five Grass Speciest ZHI-QIANG AN,' MALCOLM R

Relationships among Non-Acremonium sp. Fungal Endophytes in Five Grass Speciest ZHI-QIANG AN,' MALCOLM R. SIEGEL,1 WALTER HOLLIN,1 HUEI-FUNG TSAI,' DOROTHEA SCHMIDT,2 AND CHRISTOPHER L. SCHARDL1* Department of Plant Pathology, University ofKentucky, Lexington, Kentucky 40546-0091,1 and Federal Agricultural Research Station, CH-1260 Nyon, Switzerland2 Received 12 August 1992/Accepted 10 March 1993

Many cool-season grasses (subfamily Pooideae) possess maternally transmitted fungal symbionts which cause no known pathology and often enhance the ecological fitness and biochemical capabilities of the grass hosts. The most commonly described endophytes are the Acremonium section Albo-lanosa spp. (Acremonium endophytes), which are conidial anamorphs (strictly asexual forms) of Epichloe typhina. Other endophytes which have been noted are a Gliocladium-like fungus in perennial ryegrass (Lolium perenne L.) and a Phialophora-like fungus in tall fescue (Festuca arundinacea Schreb.). Here, we report the identification of additional non-Acremonium sp. endophytes (herein designated p-endophytes) in three more grass species: Festuca gigantea, Festuca arizonica, and Festuca pratensis. In each grass species, the p-endophyte was cosymbiotic with an Acremonium endophyte. Serological analysis and sequence determinations of variable portions of their rRNA genes indicated that the two previously identified non-Acremonium endophytes are closely related to each other and to the newly identified p-endophytes. Therefore, the p-endophytes represent a second group of widely distributed grass symbionts.

Nonpathogenic, seed-borne, fungal symbionts (endo- not been considered. Although there appear to be morpho- phytes) are a common feature of many grass species in the logical similarities (such as the penicillate conidiophores), genera Festuca and Lolium (14, 28). The most commonly these two endophytes were thought to resemble organisms observed are classified as genus Acremonium Link, section belonging to distinct genera (13). Both the Phialophora-like Albo-lanosa Morgan-Jones et Gams, and are conidial an- and Gliocladium-like endophytes were cosymbiotic with amorphs (strictly asexual forms) of the euascomycete Ep- Acremonium endophytes in their host grasses. This article ichloe typhina (Persoon: Fries) Tulasne (20). Other seed- reports the identification of similar cosymbiotic associations borne endophytes have also been described for the two grass in three additional grass species from Europe and North species Festuca arundinacea Schreb. (tall fescue) and America and presents serological and DNA sequence data Lolium perenne L. (perennial ryegrass) but have never been which indicate a close relationship between the Phialophora- formally classified. They have been designated Phialophora- like endophyte of tall fescue, the Gliocladium-like endo- like and Gliocladium-like fungi, respectively (13). phyte of perennial ryegrass, and the non-Acremonium endo- Because of their vertical transmission via seeds, grass phytes of Festuca arizonica Vasey, Festuca gigantea L., endophytes behave as heritable components of symbiotic and Festuca pratensis Huds. entities and as such, they provide important genetic, biochemical, and physiological capabilities to their hosts (24). The Acremonium endophytes are important for fitness and MATERIALS AND METHODS biological protection of their hosts (6, 24). Less is known about the ecological roles of non-Acremonium sp. endo- Biological materials. Plants possessing non-Acremonium phytes, but observations that the Phialophora-like endo- endophytes, Acremonium endophytes, or both are listed in phyte of tall fescue produces activity, in agar culture, against Table 1. Endophytes were isolated from infected grass a wide spectrum of fungal pathogens of grasses (22) suggest pseudostems by the method of Latch et al. (13). Each that it may also have a positive effect on host fitness. individual plant-fungus association is identified by a plant Extensive microscopic studies of the Gliocladium-like endo- number, and the endophyte isolates from those associations phyte in perennial ryegrass give no indication of a detrimen- are designated by the same number prefixed by either "e" tal or pathogenic effect of the endophyte upon either the for Acremonium endophytes (anamorphs of E. typhina) or grass host or the cosymbiotic endophyte, Acremonium lolii "p" for non-Acremonium endophytes (p-endophytes, Latch, Samuels, et Christensen (16-18). which, as will be shown, are related to the Phialophora-like The taxonomy and distribution of the non-Acremonium endophyte of tall fescue). E. typhina E32 was previously endophytes have received little attention. As yet, no formal identified in Festuca rubra Gaud. (21). To obtain pure binomial has been assigned to the Phialophora-like and cultures of either Acremonium or non-Acremonium endo- Gliocladium-like endophytes, and their possible relation- phytes from mixed cultures derived from plant tissues, ships to each other and to endophytes of other grasses have mycelium was ground as previously described (4); then, dilutions were plated atop a sterile cellophane disk on potato dextrose agar (PDA; GIBCO BRL, Gaithersburg, Md.). * Corresponding author. Serial 10-fold dilutions were used to obtain individual colo- t This is publication number 91-11-205 of the Kentucky Agricul- nies, which were then checked by tissue print immunoblot tural Experiment Station. assay (9) (see below). The endophyte isolates and their 1540 VOL. 59, 1993 NON-ACREMONIUM SP. GRASS ENDOPHYTES 1541

TABLE 1. Serological reactivities of grass-fungus symbiotic entities and cultures derived from them Reactivity' of pseu- Reactivity' of dostems to culture to antiserum Host Code Description antiserum Isolate(s)b against: against: p180 e19 p180 p26 e19 F. arizonica Vasey 1572 Culture from a plant in New Mexico NT NT p1572, e1572c + NT + F. arundinacea Schreb. 19 Plant artificially infected with A. co- - + e19 - - + enophialum Morgan-Jones et Gams, isolated from F. arundinacea 26 Plant artificially infected with the Phi- + - p26 + + alophora-like sp. Latch, Samuels, et Christensen isolated from F. arundinacea

F. pratensis Huds. 166-175, 179 Plants from Swiss populations natu- _d +d e166 - NT + rally infected with Acremonium un- cinatum Gams, Petrini, et Schmidt 170-178 Plants from Swiss populations +d +d e178 - NT + p178 + NT - 180 Plant from a Swiss population +d _d p180 + + F. gigantea L. 95 Breeding stock, USDA" NT NT p95, e95c + NT + p95 + NT - F. rubra commutata 32 Plant infected with E. typhina (Per- NT NT E32 - NT + Gaud. soon: Fries) Tulasne

L. perenne L. 145 Culture of Gliocladium-like sp. Latch, NT NT p145 + NT - Samuels, et Christensen (ATCC 56212) a Tissue print immunoblot. NT, not tested. b Code number of the host from which the isolate was derived, prefixed by "e' for Acremonium endophytes, "E" for E. typhina, or "p" for p-endophytes. c Mixed culture of two fungi from one symbiotic entity. d Pseudostem sections from seed progeny. USDA, U.S. Department of Agriculture, Lexington, Ky. serological reactivities are listed in Table 1. Other fungi used Toneguzzo et al. (26). Sequencing primers included those in this study are listed in Table 2. used for PCR amplification and two additional oligonucle- PCR and analysis of DNA. Fungal DNA was extracted otides, one homologous to a conserved 26S rRNA sequence from fresh or freeze-dried mycelium by published methods near position 375 (5'-GAAAAGCACTrTGAAAAGAGG (3, 4). The nuclear rRNA gene internal transcribed spacer 2 GT-3') and the other its complement (5'-ACCCTCTTTTCA (rnITS2) was amplified from genomic DNA of each isolate AAGTGCTrIT-3'). by symmetric polymerase chain reactions (PCR) using oli- Sequences of rRNA gene segments have been deposited in gonucleotide primers and conditions described by White et the EMBL data base under accession numbers X60185 al. (29). The 5' portion of the 26S large-subunit rRNA (rnL) (Acremonium coenophialum Morgan-Jones et Gams isolated was amplified by using an oligonucleotide primer homolo- from F. arundinacea), X62978 (p241), X62979 (p180), gous to a conserved 26S rRNA sequence near position 50 X62980 (pl572), X62981 (e1572), X62986 (pl78), X62987 (E. (5'-GCATATCAATAAGCGGAGGA-3') and an oligonucle- typhina from F. rubra), X62988 (Emenicella nidulans [Ei- otide primer which was complementary to a conserved dam] Vuillemin), X62989 (p26), X62990 (Gaeumannomyces segment around position 660 (5'-GACTCC1TlGGTCCGTGT graminis Arx et Oliver), and X62991 (p26). Other sequences TTCA-3'). were described by Perasso et al. (15) and Gutell et al. (8). For restriction endonuclease cleavage of symmetric PCR Phylogenetic analysis. The most parsimonious phylogram products, 10 ,ul was mixed with an equal volume of digestion was determined by using the branch-and-bound algorithm buffer (33 mM Tris-HCl [pH 7.9], 66 mM potassium acetate, implemented in PAUP version 3.0q (25). Characters were 10 mM magnesium acetate, 4 mM spermidine-HCl, 0.5 mM treated as unordered, and all nucleotide substitution differ- dithiothreitol) containing 5 U of MluI or NotI (GIBCO ences were weighted equally. Alignment gaps were consid- BRL). The mixture was incubated at 37°C for 1 to 2 h and ered equivalent to missing information. Segments that could then was analyzed electrophoretically (2). not be aligned with confidence were not included in the Single-stranded DNA templates were prepared for se- analysis. For quasi-statistical evaluation of the cladogram, quence determination by asymmetric PCR as described by bootstrap replications (7) were performed by maximum White et al. (29). Alternatively, symmetric PCR products parsimony with similar parameters and collapsing of zero- were cloned and sequenced as previously described (1). length branches. Sequences of both strands of each clone or PCR product Antiserum production and immunodetection. Antisera were determined by the Sanger method, as described by were prepared by a modification of the methods of Johnson 1542 AN ET AL. APPL. ENVIRON. MICROBIOL.

TABLE 2. Fungal cultures and their serological reactivities Reactivity' to an- Fungus Code Source tiserum against: p180 e19 Cladosporium cucumerinum Ellist et Arthur TP-20 J. Kuc, University of Kentucky - - Cochliobolus sativus (Ito et Kuribayashi) TP-17 L. Trevathan, Mississippi State Uni- - - Drechsler ex Dasture versity Colletotrichum graminicola (Ces) Wilson TP-01 N. Jackson, University of Rhode Is- - - land Drechslera erythrospila (Drechs.) Shoemaker TP-02 N. Jackson - - Drechslera siccans (Drechs.) Shoemaker TP-03 N. Jackson - - Emericella nidulans (Eidam) Vuillemin FGSC A237 Fungal Genetics Stock Center, Man- NT NT hattan, Kans. Gaeumannomyces graminis Arx et Oliver var. TP-04 N. Jackson - - avenae Turner Gaeumannomyces graminis var. graminis Arx TP-05 N. Jackson - - et Oliver Gloeotinia granigena (Quel.) T. Schumacher TP-18 S. Alderman, Oregon State University - - Laetisaria fuciformis (McAlpine) Burdsall TP-06 N. Jackson - - Leptosphaeria korrae Walke et Smith TP-07 N. Jackson - - Leptosphaerulina australis McAlpine TP-08 N. Jackson - - Limnomyces roseipellis Stalpers et Lorakker TP-09 N. Jackson - - Magnaporthe poae Landschoot et Jackson TP-10 N. Jackson - - Microdochium bolleyi (R. Sprague) De Hoog TP-11 N. Jackson - - et Hermanides-Nijhof Microdochium nivale (Fr.) Samuels et Hallett TP-12 N. Jackson - - Neurospora crassa Sheer et Dodge FGSC 2489 Fungal Genetics Stock Center NT NT Ophiosphaerella herpotrichus (Fr.: Fr.) TP-13 N. Jackson Walker Penicillium chrysogenum Thom. ATCC 9480 American Type Culture Collection + + Rhizoctonia cerealis van der Hoeven TP-15 N. Jackson Rhizoctonia solani Kuhn TP-19 K. Gwinn, University of Tennessee Sclerotinia homeocarpa Bennet TP-16 N. Jackson a Tissue print immunoblot. NT, not tested; ±, atypical reaction (see text). et al. (10) and Reddick and Collins (19). Mycelium of isolate tein extracts were dissolved in 0.5 ml of water, thoroughly e19, p26, or p180 was initially cultured for 7 to 14 days on a emulsified with 0.5 ml of Freund's incomplete adjuvant, and cellophane disk atop PDA. Subcultures were prepared by injected subcutaneously or intramuscularly. The rabbits grinding the mycelium in 10 to 15 ml of sterile water in an were injected three times at 2-week intervals and then bled 2 Omni Mixer Homogenizer (R. J. Rauch & Assoc., Holland, weeks following the last injection. Two rabbits were injected Ohio) and then inoculating 1 to 2 ml of the slurry into 50 ml with extracts from A. coenophialum e19, and four rabbits of potato dextrose broth (GIBCO BRL) in 300-ml flasks. The each were injected with extracts from p26 and p180. mycelial contents of six culture flasks were required for each Tissue print immunoblots were performed by the method rabbit to be inoculated. The cultures were grown at 21°C in of Gwinn et al. an orbital incubator (130 to 200 rpm) for 2 to 3 weeks. (9). Duplicate pseudostem sections (length, 1 Mycelia were collected by vacuum filtration and washed mm) and half-seeds were placed between a nitrocellulose with ca. 2 liters of phosphate-buffered saline minus potas- filter (pore size, 0.45 ,um; Bio-Rad, Richmond, Calif.) and sium (PBNa) (20 mM sodium phosphate buffer [pH 7.3]-150 waxed paper and then crushed onto the nitrocellulose. mM NaCl) and were resuspended in 75 ml of PBNa and Positive controls were sections of pseudostems or half-seeds homogenized in a Polytron PT20 (Brinkmann, Westbury, known to be infected with previously described endophytes N.Y.) for 45 s at setting 7. The homogenate was centrifuged (11, 12) (Table 1). Negative controls were the corresponding for 10 min at 8,000 x g. The supernatant was decanted into tissues from plants with no endophyte. Appropriate dilutions a 250-ml beaker and cooled to 4°C. Solid polyethylene glycol (typically 1:500 to 1:2,500) were determined empirically for (molecular weight, 8,000) was slowly added, with stirring, to each antiserum to obtain the most unambiguous and specific a concentration of 10% (wt/vol). The protein was precipi- identifications of positive reactions against endophytes. Col- tated overnight at 4°C, collected by a 20-min centrifugation orimetric detection of bound antibody was performed as at 12,000 x g, resuspended in 5.0 ml of PBNa containing described by Gwinn et al. (9). To avoid false-positive results, 0.1% sodium dodecyl sulfate, and then warmed to 60°C and it was important to check the color reactions on damp filters incubated for 5 min. Insoluble material was repelleted and which were well illuminated and magnified with a dissecting discarded. The supernatant was cooled on ice, and 4 vol- microscope. Positive reactions were typified by a deep red umes of cold (-20°C) acetone was added. Protein was color or deep red spots, whereas a light pink or tan color was precipitated overnight at -20°C, repelleted, resuspended in often observed with the negative controls. 5.0 ml of PBNa, and divided into 1-ml aliquots. The concen- Tissue prints of fungal mycelia were performed in similar tration of protein was estimated by A280 (2). The aliquots fashion. Duplicate samples of fresh mycelium were removed were freeze-dried for storage. For rabbit inoculations, pro- from agar plates with a 4-mm-diameter cork bore. The agar VOL. 59, 1993 NON-ACREMONIUM SP. GRASS ENDOPHYTES 1543 was cut away, and the mycelium was blotted. Each filter also TABLE 3. ELISA results for endophyte mycelia had blots of p26, p180, and e19 mycelia as controls. Indirect enzyme-linked immunosorbent assays (ELISA) OD405- utilized anti-p180 or anti-e19 antisera (diluted 1:2,000), anti- Fungus Concn A Anti-p180 rabbit (whole molecule)-alkaline phosphatase conjugate Plate 1 Plate 2 (Sigma Chemical Co., St. Louis, Mo.), and the method of Clark and Bar-Joseph (5). Optical densities at 405 nm were A. coenophialum e19 10 1.92 NT 0.38 read with a Titertek Multiskan ELISA plate reader (ICN- 1 1.39 NT 0.14 Flow Laboratories, Costa Mesa, Calif.). p180 10 NT 1.93 NT 1 NT 0.89 NT RESULTS Phialophora-like sp. p26 10 0.06 1.71 1.71 1 0.00 1.23 0.97 Identification of non-Acremonium endophytes. Microscopic examination of the Gliocladium-like endophyte in perennial P. chrysogenum 10 0.08 0.23 NT ryegrass has indicated growth characteristics which are 1 0.04 0.18 NT readily distinguishable from those of the A. lolii cosymbiont a (16). Characteristic morphologies of the Phialophora-like OD405, optical density at 405 nm; NT, not tested. endophyte and A. coenophialum in tall fescue were examined, and similar differences were observed. Leaf sheaths were prepared and examined in a fashion which is routinely against A. coenophialum e19 and the Phialophora -like p-en- and widely used to screen for endophyte infections in dophyte isolate p26 from tall fescue (Table 3). Although grasses. The Phialophora-like endophyte could be distin- these tests are not quantitative, they consistently indicated guished from A. coenophialum in stained leaf sheaths by its an intense reaction of anti-p180 antiserum against both p180 distinct hyphal morphology, as shown in Fig. 1. Similar and p26 and a weak reaction against e19. This difference was characteristics were observed for F. pratensis plants listed in most apparent at the lower concentration of mycelia (1 Table 1 which harbored non-Acremonium endophytes (data ,ug/ml). Conversely, antiserum raised against isolate e19 not shown). Also, the Phialophora-like endophyte was de- reacted most strongly with e19 and very weakly with p180 tected in tall fescue seed endosperm, though mycelia were and p26. Neither of these two antisera showed appreciable sparse (Fig. 1). Later tissue print immunoblot assays of reaction to another ascomycetous fungus, Penicillium half-seeds suggested that the hyphae of this non-Acremo- chrysogenum Thom. nium endophyte were mainly concentrated in the embryo The specificities of the antisera for either p-endophytes or (data not shown). Acremonium endophytes allowed discrimination of the en- The non-Acremonium endophytes were isolated from sur- dophyte types in tissue print immunoblot assays (Table 1). face sterilized pseudostems of seed progeny from F. praten- Non-Acremonium p-endophytes p180 and p26 both reacted sis plants 176, 178, and 180. They were typically slow- with anti-p180 and anti-p26 but not with anti-e19. Con- growing, white in color, and difficult to distinguish from versely, anti-e19 showed reactivity againstA. coenophialum Acremonium endophyte cultures at the macroscopic level. e19 but not against p180 or p26. Antiserum from each of the They were found to be distinct from Acremonium endo- inoculated rabbits showed specificity for either p-endophytes by serology (tissue print immunoblot and ELISA) phytes or Acremonium endophytes strictly in accordance and by rrnITS2 sequence (see below). The non-Acremonium with the type of endophyte with which the rabbit had been endophyte in progeny of plant 180 was consistently obtained inoculated (data not shown). in pure culture, but those from progeny of plants 176 and 178 Pure cultures of p95 and p178 and the Gliocladium-like were isolated as mixed cultures with Acremonium endo- endophyte from perennial ryegrass were also tested by tissue phytes. By plating serial dilutions of ground mycelia, several print immunoblot with both anti-p180 and anti-e19 (Table 1). non-Acremonium endophytes which exhibited no significant They all reacted strongly with the anti-p180 antiserum but reaction to antiserum against A. coenophialum were ob- not with anti-e19. Conversely, pure cultures ofAcremonium tained in pure culture. Non-Acremonium endophytes were endophytes e166 and e178 from F. pratensis reacted with also obtained from F. arizonica plant 1572 and F. gigantea anti-e19 but not with anti-p180. plant 95 (Table 1), and in both cases, Acremonium cosym- The antisera were also sufficiently specific to discriminate bionts were also present. Together with the previous identi- the two endophyte types in pseudostem sections from F. fications of endophytes termed Phialophora-like and Glio- pratensis and F. arundinacea plants (Table 1). Those plants cladium-like in F. arundinacea and L. perenne, respectively from which mixed cultures had been obtained reacted with (13), this amounted to five plant species in which cosymbi- both antisera, those from which pure p-endophyte cultures oses ofAcremonium and non-Acremonium endophytes were were obtained reacted with anti-p180 but not anti-e19, and observed. plants from which pure Acremonium cultures were obtained Serological relationships of non-Acremonium p-endophytes. reacted with anti-e19 but not with anti-p180. Tall fescue Serological analyses indicated a relationship between the plants 26 and 19, which had specific endophytes introduced non-Acremonium endophytes of all five grass species and into them (p26 and e19, respectively) reacted as expected. In also suggested that they were not closely related toAcremo- addition, 30 half-seeds from plant p26 were analyzed by the nium endophytes. Because of this relationship, the non- same method; 86% were serologically positive for p-endo- Acremonium endophytes identified in these five grass spe- phytes, and as expected, all were negative for Acremonium cies are hereafter referred to collectively as p-endophytes. endophytes. In contrast, seeds from F. arundinacea cv. Antiserum was raised against isolate p180, the p-endophyte Kentucky 31 were consistently positive for Acremonium from seed progeny of F. pratensis plant 180. The reactivity endophytes and negative for p-endophytes. of the anti-p180 antiserum was tested by indirect ELISA Maternal-line transmission of p-endophytes and Acremo- 1544 AN ET AL. APPL. ENVIRON. MICROBIOL.

FIG. 1. Photomicrographs of p-endophytes (a and c) and A. coenophialum (b and d) in the seed aleurone tissue (a and b) and leaf sheaths (c and d) of F. arundinacea, stained by the method of Latch et al. (14). Fungal hyphae are indicated by arrows. Note that, in seed tissue (a), the p-endophyte is highly septated. The A. coenophialum mycelium (b) is much more densely packed around the darkly stained aleurone cells. In leaf sheaths (c), the p-endophyte produces straight, branched, intercellular hyphae, which are thinner than those ofA. coenophialum (d). The latter has highly convoluted hyphae which typically follow the longitudinal axis of the host cells. Bar, 10 p.m (magnification, x400). VOL. 59, 1993 NON-ACREMONIUM SP. GRASS ENDOPHYTES 1545

rrn5.8S p26 (Festuca arundinacea) TAATGTGAATTGCAGAATTCAGTGAATCATCGAATCTTTGAACGCATATTGCGCCCTCTGGTATTCCGGGGGGCATGCCTGTTCGAGCGTCATTA

p1572 (Festuca arizonica) ...... p95 (Festuca gigantea) ......

p178 (Festuca pratensis) ......

p180 (F. pratensis) ......

p241 (Lolium perenne) ......

rrnITS2 p26 (F. arundinacea) TAACCACTCAAGCTCTCGCTTGGTATTGGGGTTCGCGGTCTCGCGGCCCCTAAAATCAGTGGCGGTGCCTGTCGGCTCCGCG GTAATACT

p1572 (F. arizonica) ...... p95 (F. gigantea) p178 (F. pratensis) p180 (F. pratensis) p241 (L. perenne)

rrnL p26 (F. arundinacea) CCTCGCGTCTGGGTCCGACAGGTCTACTTGCCAACAACCCCCAATTTTTTACAGGTTGACCTCGGATCAGGTAAGGGATACCCGCTGAACTTAA

p1572 (F. arizonica) ...... AA.. p95 (F. gigantea) p178 (F. pratensis)

p180 (F. pratensis) ......

p241 (L. perenne) ...... A ...... AA.- FIG. 2. Sequence alignments of a segment of internal transcribed spacer 2 (rmITS2) and flanking portions of rRNA structural genes from the indicated p-endophytes. Identity to the reference sequence at the top is indicated by dots; alignment gaps (insertion or deletion differences) are indicated by dashes. The MluI site common to the p-endophyte rrnITS2 sequences is boxed. nium endophytes in F. pratensis was confirmed by tissue that between the rrnITS2 sequences of Acremonium endo- print immunoblots. Of five progeny of plants 176 and 177, all phytes and those of E. typhina isolates, which differed by up were positive for both p-endophytes and Acremonium endo- to 16 nucleotide substitutions (1, 20). Thus, the Phialophora- phytes. Of five progeny of plant 178, four were positive for like endophyte, the Gliocladium-like endophyte, and other p-endophytes and all were positive for Acremonium endo- p-endophytes exhibited sequence similarities suggestive of phytes. None of the progeny of plant 180 appeared to have congeneric or even conspecific relationships. Acremonium endophytes, but four of the five were positive Inspection of the rnITS2-sequence indicated the possibil- for p-endophytes. Seeds from other plants gave rise to ity that PCR, followed by restriction endonuclease cleavage, progeny which consistently possessed Acremonium endo- could be a rapid, nonserological method to distinguish phytes. Three to five progeny each from plants 166 to 175 p-endophytes from Acremonium endophytes. The p-endo- and 179 were screened, and all but one (from plant 169) were phyte rrnITS2 fragments were cleaved with MluI but not positive for Acremonium endophytes and negative for p-en- with NotI, whereas Acremonium endophyte rmITS2 con- dophytes. tained a NotI site but no MluI site. Thus, amplification and In order to further assess the usefulness of the antisera in cleavage of this DNA segment resulted in a pattern charac- distinguishing the endophytes from several possible grass teristic of the type of endophyte from which the DNA was pathogens and saprophytes which may occur in seed or obtained (Fig. 4). Both patterns were obtained from mixed vegetative tissues, the serological reactivities of 20 other cultures (data not shown). fungi were tested by tissue print immunoblot (Table 2). In To investigate the relationship between the p-endophytes each case, mycelia from p180 and e19 were included as and euascomycetes, particularly E. typhina and theAcremo- controls. No strong positive reactions against any non- nium endophytes, sequence comparisons of more conserved endophytic fungi were observed. Only P. chrysogenum gave portions of the rRNA genes were made (Fig. 3). Sequences an apparent weak reaction, but examination under a dissect- which could not be aligned unambiguously (lowercase letters ing microscope indicated that the color was of a light shade in Fig. 3) were omitted from the phylogenetic analysis. The (pink rather than deep red), as was sometimes also observed with the negative controls and uninfected plant tissues. A remaining regions exhibited enough usable variation for The E. typhina and A. coenophialum similar specificity was observed for the anti-e19 antiserum in parsimony analysis. were identical in the and were, tests in which e19 and p26 mycelia served as controls. Again, sequences aligned regions as a taxonomic unit. The most among the nonendophytic fungi listed in Table 2, only P. therefore, treated single single tall fescue chrysogenum gave a reaction that may have been interpreted parsimonious tree (Fig. 5) placed the p-endophyte to the as positive, but on microscopic examination it was judged (p26) within the euascomycetes but closer plecto- atypical of the positive control. mycetes than to the pyrenomycetes. In contrast, A. coeno- DNA sequence relationships of the p-endophytes. The pos- phialum was placed with the pyrenomycetes (represented by sible relationships of the p-endophytes with each other and E. typhina, G. graminis, and Neurospora crassa Sheer et with the Acremonium endophytes were investigated by Dodge). Bootstrap replications suggested a high confidence sequence analysis of rrnITS2 (Fig. 2 and 3). The p-endophyte for separating the pyrenomycete clade from the other fungi, sequences were closely related to each other, differing by 0 including the p-endophyte, E. nidulans, and P. chrysoge- to 3 nucleotide substitutions and 1 to 4 positions with num. Although these results do not identify the specific alignment gaps, in a total of 148 positions (Fig. 2). Although taxonomic position of the p-endophytes within the euasco- none of the sequences were identical, the similarity between mycetes, they do indicate distinct evolutionary origins of the the p-endophyte rrnITS2 sequences was much greater than two endophyte types. 1546 AN ET AL. APPL. ENVIRON. MICROBIOL.

. 140 . 150. Acremoni um coenophi alum, rrn5.8S ...... 0...... 90 100 . 110 . 120 . 130 Epichloe typhi.na TAATGTGAATTGCAGAATTCAGTGAATCATCGAATCTTTGAACGCACATTGCGCCCGCCQGTATT-CTGGCGGGCATGCCTGTTCGAGCGTCATTT p-endophyte p26 TAATGTGAATTGCAGAATTCAGTGAATCATCGAATCTTTGAACGCATATTGCGCCCTCTGGTATT-CCGGGGGGCATGCCTGTTCGAGCGTCATTA Saccharomyces cerevi si ae TAAGGTGAATTGCAGAATTCCGTGAATCATCGAATCTTTGAACGCAQTTGCGCCCCTTGGTATT-CCAGGGGGCATGCCTGTTTGAGCGTCATTT Emericella nidulans TAATGTGAATTGCAGAATTCAGTGAATCATCGAGTCTTTGAACGCACATTGCGCCCCCTGGCATT-CCGGGGGGCATGCCTGTCCGAGCGTCATTG Penicillium chrysogenum TAATGTGAATT-CGAAATTCAGTGAATCATCGAGTCTTTGAACGCAQTTGCGCCCCCTGGTATT-CCGGGCGGCATGCCTGTCCGAGCGTCATTT Gaeumannomyces gramini s TAATGTGAATTGCAGAATTCAGTGAATCATCGAATCTTTGAACGCAQTTGCGNNNNNCGGTATT-CCGGTGGGCATGCCTGTCCGAGCGTCATTT Neurospora crassa TAATGTGAATTGCAGAATTCAGTGAATCATCGAATCTTTGAACGCACATTGCGCTCGCCAGTATT-CTGGCGAGCATGCCTGTTCGAGCGTCATTT * * * ** * * * * rrnITS2.10 ...... 20 .... 30 .... 40.... 50 .... 60 ...... 70 .... 80 .... 90 A. coenophialum caacc-ctcaagcccgctgcgcgcttgc-tgttggggaccggctcacccgcctcgcggcgbCgccCkcecgaaatgaatcggcggtct p-endophyte p26 taaccactcaagctct-----cgcttg-gtattgggg.------ttcgcggtctcgcggcccctaaaatc Emericella nidulans ctgcc-ctcaagcccg.------gcttgtgtgttggg.------.tcgtcgtcccccccggggacgggCccgaaaggcagcggcgcac Gaeumannomyces gramini s c-accactcaagccca------gcttg-gtgttgggg---ctcccggcgcccggcggtcggggCccccaagtacatcggcggtctcgcca

. . 130 . 140 . 150 . 160 . 170 . 180 ...... 100 . 110 120 A. coenophialum cgtcgcaagcctcctttgcgtagtagcacaccacctcgcaaccgggagcgcggcgcggccactgccgtaaaacgcccaactttctccaaga p-endophyte p26 agtggcggtgcctgtcggctct c c t gtaatactcctcgcgtctgggtccgacaggtctacttgccaacaacccccaattttttacag Emericella nidulans cgtgtccggtcctgagcgtatggggctttgtcacccgctcgattagggccggccgggcgccagccggcgtctccaaccttatttttctcag Gaeumannomyces gram n s ggaccctgaacgcagtaactcgcggtaaaacgcgacttcgttcggaggcttcccggcaaactccagccgctaaaccccctaaacttcttag

rrnL..... 10 . 20 . 30 . 40 . 50 . 60 . 70 ...... 80 90 . 100 A. coenophialum, E. typhina GTT-GACCTCGAATCAGGTA-GGACTACCCGCCGAACTTAAGCATATCAATAAGCGGAGGAAAAGAAACCAACAGGGATTGCCCCAGTAACGGCGAGTGA p-endophyte p26 GTT-GACCTCGGATCAGGTAAGGGATACCCGCTGAACTTAANNNNNNNNNNNNNNNNNNNNAAAGAAACCAACAGGGATTACCTCAGTAACGGCGAGTGA Saccharomyces cerevi si ae GTTTGACCTCAAATQGGTA-GGAGTACCCGCTGAACTTAAGCATATCAATAAGCGGAGGAAAAGAAACCAACCGGGATTGCCTTAGTAACGGCGAGTGA Emericella nidulans GTT-GACCTCGGATCAGGTA-GGGATACCCGCTGAACTTAANNNNNNNNNNNNNNNNNNNNAAAGAAACCAACAGGGATTGCCTCAGTAACGGCGAGTGA Penicillium chrysogenum GTT-GACCTCGGATCAGGTA-GGGATACCCGCTGAACTTAANNNNNNNNNNNNNNNNNNNNAAAGAAACCAACAGGGATTGCCCCAGTAACGGCGAGTGA Gaeumannomyces graminis GTT--GCCTCGGATCAGGTA-GGAATACCCGCTGAACTTAANNNNNNNNNNNNNNNNNNNNAAAGAAACCAACAGGGATTGCCCTAGTAACGGCGAGTGA Neurospora crassa GTT-GACCTCGGATCAGGTA-GGAATACCCGCTGAACTTAAGCATATCAATAAGCGGAGGAAAAGAAACCAACAGGGATTGCCCTAGTAACGGCGAGTGA

. . 150 . 160 . 170 ...... 180 . 190 . 200 ...... 110 . 120 130 140 . A. coenophialum, E. typhina AGCGGCA-ACAGCTCAAATTTGAAATCtggccC-----cggggcccGAGTTGTAATTTGCAGAGGAtgctttt-ggcgaggcgCct------p-endophyte p26 AGCGGTA-ACAGCTCAAATTTGAAAGCtagctcttt.---tagggctcGCATTGTAGTTTGTAGAAGAtgcttcg-agtgtgg ccC------g Saccharomyces cerevisiae AGCGGCA-AAAGCTCAAATTTGAAATCtggtacctt----cggtgcccGAGTTGTAATTTGGAGAGGGcaaCttt-ggggccgttcCtt------Emericella nidulans AGCGGCA-ACAGCTCAAATTTGAAAGCtggcccctt.--cggggtccGCGTTGTAATTTGCAGAGGAtgcttcg-ggtgcggcccct------Penicillium chrysogenum AGCGGCA-AGAGCTCAAATTTGAAAGCtggctctt ---cggggtccGCATTGTAATTTGTAGAGGAtgcttcg-ggtgcggtCCcc.------Gaeumannomyces graminis AGCGGCA-ACAGCTCAAATTTGAAATCtggccc .-----.taggcccGAGTTGTAATTTGCAGAGGAtgctttt-ggcaaagcgctt------Neurospora crassa AGCGGCA-ACAGCTCAAATTTGAAATCtgge------ttcggcccGAGTTGTAATTTGTAGAGGAagCtttt-ggtgaggcacct------** *

. . 250 . 260 . 270 ...... 280 . 290 . 300 ...... 210 . 220 . 230 240 A. coenophialum, E. typhina -TCCGAGTTCCCTGGAACGGGACGC----CACAGAGGGTGAGAGCCCCGTct---ggttggatgcCgagCctc-tGTAAAGCTcCTTCGACGAGTCGAGT p-endophyte p26 gTCTAAGTTTCTTGGAACAGGACGT----CATAGAGGGTGAGAATCCCGTatg-tgactgggtgccttcgctcacGTGAAGCTCTTTGACGAAGTCGAGT Saccharomyces cerevisiae gTCTATGTTCCTTGGAACAGGACGT----CATAGAGGGTGAGCATCCCGTgtggcgagga---gtgcggttctttGTAAAGTGCCTTCGAAGAGTCGAGT Emericella nidulans gTCTAAGTGCCCTGGAACGGGCCGT----CAGAGAGGGTGAGAATCCCGTct-tgggcagggtgcc-gtgcccgtGTGAAGCTCCTTCGACGAGTCGAGT Penicillium chrysogenum aTCTAAGTGCCCGTGAGGGACGT--TAGAGTGAGAATCCCGTat-gggatggggtgtccgcgcccgtGTGAAGCTCCTTCGAGAGTCGAGT Gaeumannomyces gramini s -ACCGAGTCCCCTGGAACGGGGCGC----CACAGAGGGTGAGAGCCCCGTat--ggtatgac--gccgagcctctGTAAAGCTCCTTCGACGAGTCGAGT Neurospora crassa -TCTGAGTCCCCTGGAACGGGGCGC----CATAGAGGGTGAGAGCCCCGTat--agtc-ggc-tgccgatccaatGTAAAGCTCCTTCGACGAGTCGAGT

...... 400 ...... 380 390 .. . 370 ...... 3 0...... 1320 .. . . 330 . 340 350 360 A. coenophialum, E. typhina AGTTTGGGAATGCTGCTCTAAATGGGAGGTATATTTCTTCTAAAGCTAAATATTGGCCAGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA p-endophyte p26 TGTTTGGGAATGCAGCTCAAAATGGGTGGTAAATTTCATCTAAAGCTAAATATTGGCCAGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA Saccharomyces cerevisiae TGTTTGGGAATGCAGCTCTAAGTGGGTGGTAAATTCCATCTAAAGCTAAATATTGGCGAGAGACCGATAGCGAACAAGTACAGTGATGGAAAGATGAAAA Emericella nidulans TGTTTGGGAATGCAGCTCTAAATGGGAGGTAAATTTCATCTAAAGCTAAATACCGGCCGGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA Peni cillium chzysogenum TGTTTGGGAATGCAGCTCTAAATGGGTGGTAAATTTCATCTAAAGCTAAATATTGGCCGGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA Gaeumannomyces graminis AGTTTGGGAATGCTGCTCTAAATGGGAGGTAAATTTCTTCTAAAGCTAAATACCGGCCAGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA Neurospora crassa AGTTTGGGAATGCTGCTCAAAATGGGAGGTAAATTTCTTCTAAAGCTAAATATTGGCCAGAGACCGATAGCGCACAAGTAGAGTGATCGAAAGATGAAAA * * * * * **

0 . . 420 . 430 . 440 . 450 . 460 ...... 41 A. coenophialum, E. typhina GCACTTTGAAAAGAGGGTTAAATAG-TACGT-GAAATTGTTGAAAGGGAAGCGCTCATGACCAGACT p-endophyte p26 GCACTTTGGAAAGAGAGTTAAACAGCTACCTTGAAATTGTTGAAAGGGAAGCGCTTGCAACCAGACC Saccharomyces cerevisiae GAACTTTGAAAAGAGAGTGAAAAAG-TACGT-GAAATTGTTGAAAGGGAAGGGCATTTGATCAGACA Emericella nidulans GCACTTTGAAAAGAGAGTTAAACAGCCACCT-GAAATTGTTGAAAGGGAAGCGCTTGCGACCAGACT Penicillium chrysogenum GCACTTTGAAAAGAGAGTTAAAAAG-CACGT-GAAATTGTTGAAAGGGAAGCGCTTGCGACCAGACT Gaeumannomyces gramini s GCACTTTGAAAAGAGGGTTAAATAG-CACGT-GAAATTGTTGAAAGGGAAGCGCTCGTGACCAGACT Neurospora crassa GCACTTTGAAAAGAGGGTTAAATAG-CACGT-GAAATTGTTGAAAGGGAAGCGTTTGTGACCAGACT * * * * * * FIG. 3. Sequence alignments of a segment of the rRNA gene repeats encoding the 3' portion of the 5.8S rRNA (nm5.8), internal transcribed spacer 2 (rrnITS2), and the 5'-terminal portion of the 26S large-subunit rRNA (rmL). Positions in the rmL segment were aligned and numbered according to the system of Perasso et al. (15). Regions which were too divergent for unambiguous alignment and, therefore, were not used in phylogenetic analysis are indicated by lowercase letters; positions that were informative and used for parsimony (Fig. 5) are indicated by asterisks. The NotI site in rrnITS2 of A. coenophialum and the MluI site in rrnITS2 of p-endophyte p26, which were used for restriction analysis of PCR products, are boxed. VOL. 59, 1993 NON-ACREMONIUM SP. GRASS ENDOPHYTES 1547

1 2 a A -9 rapidly disseminated than the Acremonium endophytes and E. typhina. The p-endophyte of perennial ryegrass sporu- lates in nature (16), and unlike theAcremonium endophytes, it can be introduced into host seedlings without prior wound- ing of the plant (12). In contrast, Acremonium endophytes are exclusively transmitted via host seeds, rhizomes, and tillers (23). Spore dissemination may lead to a more rapid spread of the p-endophytes geographically and among different hosts. For this reason, it is possible that the p-endophytes in this study have a more recent common ancestry than do the Acremonium endophytes from the same host FIG. 4. Analysis of amplified DNA spanning rrnITS2 by restric- grasses. tion endonuclease cleavage and electrophoresis. Lane 2, the ampli- In addition to taxonomy and phylogeny, there are other fied product from A. coenophialum e19, subsequently cleaved with interesting differences between p-endophytes and Acremo- NotI; lane 3, the e19 product treated with MluI, which did not cleave nium endophytes. The most thoroughly studied p-endophyte it; lane 4, the amplified product from p-endophyte p26, treated with is that of perennial ryegrass, which (like the tall fescue NotI and remaining uncleaved; lane 5, the p26 product cleaved with p-endophyte) is found in virtually all plant tissues (13, 16). In MluI. The sizes (in base pairs) of DNA markers (lane 1) are indicated on the left. contrast, Acremonium endophytes are not observed in roots and can have restricted distribution in leaf blades (23). Also, the L. perenne p-endophyte has been observed to invade DISCUSSION host cells and to sporulate, but only in senescent host tissues (18). For this reason, the p-endophyte produced spores The results reported herein indicate that the non-Acremo- without causing any apparent disease, unlike E. typhina, nium p-endophytes can occur cosymbiotically with Acremo- which sterilizes host panicles during its sporogenous stage. nium endophytes in several species of the related grass In contrast to the Acremonium endophytes, the possible genera Festuca and Lolium. They further indicate that the benefits of the p-endophytes remain relatively unexplored. p-endophytes, including those previously distinguished as The wide spectrum of antifungal activity exhibited in agar Phialophora-like and Gliocladium-like (13), are closely re- culture by the p-endophyte of tall fescue suggests that it may lated. Thus, they constitute a second group of endophytes, enhance resistance to fungal diseases (22). Since p-endo- which is more widely distributed than previously suspected. phytes often occurred together with Acremonium endo- Serological and PCR techniques were adapted for their phytes in Festuca and Lolium spp., it is also possible that the identification in host vegetative tissues and seeds and in two mycosymbionts have synergistic activities in biological culture. protection and other aspects of host fitness. However, the Although the two p-endophytes previously reported were possibility that some or all of the p-endophytes are commen- not originally considered congeneric (13), their rnITS2 sal symbionts or are even antagonistic to their hosts must sequence homology is actually much higher than that exhib- also be considered. Their apparent antifungal activities and ited by different isolates of the single teleomorphic species the lack of obvious disease symptoms associated with p-en- E. typhina and much higher than that between Acremonium dophyte infection do not constitute adequate evidence of endophytes of the genera Festuca and Lolium (20). Two mutualism. Extensive fitness studies comparing two-part possibilities, which are not mutually exclusive, may explain symbiotic entities (plants with one endophyte) and three-part this relatively close sequence similarity of the p-endophytes. symbiotic entities (plants with cosymbiotic Acremonium First, rates of evolution of the rrnITS2 sequences in these endophytes and p-endophytes) with host fungi are not known, and it is possible that the Acremonium grasses lacking endophytic fungi need to be undertaken. If the sequences evolve faster than those of the p-endophytes. p-endophytes are found to be or Second, it is also possible that the p-endophytes are more deleterious to their hosts otherwise undesirable (e.g., toxic to livestock), then the serological and PCR-based detection techniques described here may be E. typhina, used to screen seed lots and thereby reduce their presence in A. coenophialum commercial cultivars. If the p-endophytes are mutualists, G. then these tools can be used in a program to introduce them graminis into cultivars. The ability of Festuca and Lolium spp. to harbor non- p-endophyte pathogenic, seed-borne mycosymbionts effectively increases their biological complexity and heritable genetic diversity. Em. nidulans Under certain conditions of biotic or abiotic stress, these P. chrysogenum grasses appear to be ecologically dependent upon their Acremonium endophytes (6, 23, 27). However, it is possible S. cerevisiae that, in many tests of the benefits of Acremonium endo- FIG. 5. The most parsimonious gene tree based on aligned 5.8S phytes, it was unknown whether the Acremonium endo- and rrnL sequences shown in Fig. 3. A distance of 10 nucleotide phytes were present alone or in cosymbiosis with p-endo- substitution differences is indicated by a bar at the upper left. The phytes. In the future, it will be important to determine the frequencies with which each branch was represented in 1,000 status of experimental with to the of bootstrap replications (7) are indicated. The total length of each tree plants regard presence was 122, and the consistency index, excluding autapomorphies, was both endophyte types and to investigate the ecological 0.710. The root of the tree was established by using orthologous importance of the p-endophyte-grass interactions as well as sequences from Mus musculus and L. perenne. the possible interactions of cosymbiotic endophytes. 1548 AN ET AL. APPL. ENvIRON. MICROBIOL.

ACKNOWLEDGMENTS species. Ann. Appl. Biol. 111:59-64. 15. Perasso, R., A. Baroin, L. H. Qu, J. P. Bachellerie, and A. We thank Alfred D. Byrd and Ulla Jarlfors for technical assistance Adoutte. 1989. Origin of the algae. Nature (London) 339:142- and James F. White, Jr., for helpful discussions. 144. This work was supported by the McKnight Foundation, St. Paul, 16. Philipson, M. N. 1989. A symptomless endophyte of ryegrass Minn., and by U.S. Department of Agriculture grants CRGO (Lolium perenne) that spores on its host-a light microscope 88-37151-3860 and NRICGP 92-00694. study. N.Z. J. Bot. 27:513-519. 17. Philipson, M. N. 1991. Ultrastructure of the Gliocladium-like REFERENCES endophyte of perennial ryegrass (Lolium perenne L.). I. Vege- 1. An, Z.-Q., J.-S. Liu, M. R. Siegel, G. Bunge, and C. L. Schardl. tative phase and leaf blade sporulation. New Phytol. 117:271- 1992. Diversity and origins of endophytic fungal symbionts of 280. the North American grass Festuca arizonica. Theor. Appl. 18. Philipson, M. N. 1991. Ultrastructure of the Gliocladium-like Genet. 85:366-371. endophyte of perennial ryegrass (Loliumperenne L.). II. Sporu- 2. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. lation in the leaf sheath. New Phytol. 117:281-288. Seidman, J. A. Smith, and K. Struhl (ed.). 1992. Current 19. Reddick, B. B., and M. H. Collins. 1988. An improved method protocols in molecular biology. John Wiley & Sons, Inc., New for detection ofAcremonium coenophialum in tall fescue plants. York. Phytopathology 78:418420. 3. Bruns, T. D., R. Fogel, and J. W. Taylor. 1990. Amplification 20. Schardl, C. L., J.-S. Liu, J. F. White, R. A. Finkel, Z. An, and and sequencing of DNA from fungal herbarium specimens. M. R. Siegel. 1991. Molecular phylogenetic relationships of Mycologia 82:175-184. nonpathogenic grass mycosymbionts and clavicipitaceous plant 4. Byrd, A. D., C. L. Schardl, P. J. Songlin, K. L. Mogen, and pathogens. Plant Syst. Evol. 178:2741. M. R. Siegel. 1990. The 0-tubulin gene of Epichloe typhina from 21. Schardl, C. L., and H.-F. Tsai. 1992. Molecular biology and perennial ryegrass (Lolium perenne). Curr. Genet. 18:347-354. evolution of the grass endophytes. Nat. Toxins 1:171-184. 5. Clark, M. F., and M. Bar-Joseph. 1984. Enzyme immunosor- 22. Siegel, M. R., and G. C. M. Latch. 1991. Expression of bent assays in plant virology. Methods Virol. 7:51-85. antifungal activity in agar culture by isolates of grass endo- 6. Clay, K. 1990. Fungal endophytes of grasses. Annu. Rev. Ecol. phytes. Mycologia 83:525-537. Syst. 21:275-295. 23. Siegel, M. R., G. C. M. Latch, and M. C. Johnson. 1987. Fungal 7. Felsenstein, J. 1985. Confidence limits on phylogenies: an ap- endophytes of grasses. Annu. Rev. Phytopathol. 25:293-315. proach using the bootstrap. Evolution 39:783-791. 24. Siegel, M. R., and C. L. Schardl. 1991. Fungal endophytes of 8. Gutell, R. R., M. N. Schnare, and M. W. Gray. 1990. A grasses: detrimental and beneficial associations, p. 198-221. In compilation of large subunit (23S-like) ribosomal RNA se- J. H. Andrew and S. S. Hirano (ed.), Microbial ecology of quences presented in a secondary structure format. Nucleic leaves-1991. Springer-Verlag, Berlin. Acids Res. 18:2319-2330. 25. Swofford, D. L. 1991. PAUP: phylogenetic analysis using parsi- 9. Gwinn, K. D., M. H. Shepard-Collins, and B. B. Reddiclk 1991. mony, version 3.0q. Illinois Natural History Survey, Cham- Tissue print-immunoblot: an accurate method for the detection paign. of Acremonium coenophialum in tall fescue. Phytopathology 26. Toneguzzo, F., S. Glynn, E. Levi, S. Mholsness, and A. Hayday. 81:747-748. 1988. Use of a chemically modified T7 DNA polymerase for 10. Johnson, M. C., T. P. Pirone, M. R. Siegel, and D. R. Varney. manual and automated sequencing of supercoiled DNA. Bio- 1982. Detection of Epichloe typhina in tall fescue by means of Techniques 6:460469. enzyme-linked immunosorbent assay. Phytopathology 72:647- 27. White, J. F. 1988. Endophyte-host associations in forage 650. grasses. XI. A proposal concerning origin and evolution. My- 11. Johnson, M. J., M. R. Siegel, and B. A. Schmidt. 1985. Serolog- cologia 80:442446. ical reactivities of endophytic fungi from tall fescue and peren- 28. White, J. F., Jr., and G. T. Cole. 1985. Endophyte-host associ- nial ryegrass and of Epichloe typhina. Plant Dis. 69:200-202. ations in forage grasses. I. Distribution of fungal endophytes in 12. Latch, G. C. M., and M. J. Christensen. 1985. Artificial infec- some species of Lolium and Festuca. Mycologia 77:323-327. tions of grasses with endophytes. Ann. Appl. Biol. 107:17-24. 29. White, T. J., T. D. Bruns, S. Lee, and J. Taylor. 1990. Analysis 13. Latch, G. C. M., M. J. Christensen, and G. J. Samuels. 1984. of phylogenetic relationships by amplification and direct se- Five endophytes of Lolium and Festuca in New Zealand. quencing of ribosomal RNA genes, p. 315-322. In M. A. Innis, Mycotaxon 20:535-550. D. H. Gelfand, J. J. Sninsky, and T. J. White (ed.), PCR 14. Latch, G. C. M., L. M. Potter, and B. F. Tyler. 1987. Incidence protocols: a guide to methods and applications-1990. Aca- of endophytes in seeds from collections of Lolium and Festuca demic Press, San Diego, Calif.